How to 480i to 720p

MikeLV wrote on 3/8/2020, 5:29 PM

Have some older footage I need to convert to 720P for Vimeo viewing, with the following data:

ID                                       : 0
Format                                   : DV
Commercial name                          : DVCPRO
Duration                                 : 39 min 33 s
Bit rate mode                            : Constant
Bit rate                                 : 24.4 Mb/s
Width                                    : 720 pixels
Height                                   : 480 pixels
Display aspect ratio                     : 4:3
Frame rate mode                          : Constant
Frame rate                               : 29.970 (29970/1000) FPS
Standard                                 : NTSC
Color space                              : YUV
Chroma subsampling                       : 4:1:1
Bit depth                                : 8 bits
Scan type                                : Interlaced
Scan order                               : Bottom Field First
Compression mode                         : Lossy
Bits/(Pixel*Frame)                       : 2.357
Time code of first frame                 : 00:39:33;25
Time code source                         : Subcode time code
Stream size                              : 7.95 GiB (100%)
Encoding settings                        : wb mode= / white balance= / fcm=manual focus

What would be the correct steps to follow? I'll pan/crop to 16:9 to avoid black bars on the sides. I have HOS installed if that would be a better choice at render time. But unsure about how to set the project properties, and render settings. Thanks for any suggestions.

Comments

Musicvid wrote on 3/8/2020, 5:46 PM

HOS with Lanczos resize and it's Deinterlacer (QTGMC) is your best bet.

MikeLV wrote on 3/8/2020, 6:13 PM

Thank you, that actually looks pretty good! According to the qtgmc wiki, the default is "slower" Is that what you would normally use?

Musicvid wrote on 3/8/2020, 7:45 PM

It's really slow anyway, and I can't tell any difference at Medium.

NickHope wrote on 3/8/2020, 11:37 PM

Here's an AviSynth script you can use for this with HOS AviDub. Lanczos3 and Spline36 resizing give a similar result. I would render to MagicYUV RGB and bring the 1280 x 960 back into Vegas for cropping and sharpening.

SetMemoryMax(16384)
AviSource("d:\fs.avi")
AssumeBFF()
ConvertToYV12(interlaced=true)
QTGMC(Preset="Slower", EdiThreads=4)
Spline36Resize(1280,960)
AssumeFPS(59.94)
Prefetch(16)

For performance/stability, you might need to reduce some of the numbers depending on your rig. For example if it crashes, try halving the 16384, 4 and 16.

However don't expect too much. It's basically impossible to get a sharp result when doubling the number of lines, no matter what fancy deinterlacing/resizing/sharpening you use. If 16:9 isn't essential, you might consider a more modest resize to 960 x 720, which is still "HD" but 4:3 of course. YouTube accept such resolution without you rendering black side-bars to make it 16:9, so I guess Vimeo would too. Or you could consider a hybrid of cropping and modest horizontal stretching. That's worked well for me before, although the subject was fish and not humans.

MikeLV wrote on 3/9/2020, 4:59 PM

@NickHope, wouldn't rendering twice cause quality loss? I try to avoid having to render something twice. I don't care if it goes slow, I'll go do other things while it renders. Is there some benefit to using the script instead of the settings available in HOS? I ran a test render of a minute of footage and it looked pretty good just doing what musicvid said.

NickHope wrote on 3/10/2020, 1:30 AM

@MikeLV Rendering twice in this case is not a problem because MagicYUV RGB is lossless.

RenderPlus does fundamentally give as good quality as AviDub. However, two reasons I'm used to doing a lossless intermediate before render:

1. I like to add titles/graphics at the final render resolution, for maximum quality. If you add these to the original 480i project, they will get a bit mangled during processing. If you don't have titles/graphics then no problem.

2. I've never really been happy with AviSynth sharpening, and that goes for its basic "Sharpen" as well as the fancier "LimitedFasterSharpen" (used by RenderPlus) and "LSFMod". It's just not that sharp! Also, I personally think sharpening should come after resizing, but RenderPlus puts it before resizing. So I like to add some final native Vegas sharpening to my lossless intermediates at the final render resolution. The danger with the Vegas sharpening is that AFAIK it makes no effort to identify the edges of objects, so it boosts noise. So you have to be careful not to use too much. But the benefit is that it's more crisp than AviSynth sharpening. In either case, for your project, you would probably want to do some quite aggressive sharpening. e.g. RenderPlus sharpen of at least 500, or VEGAS Sharpen of say 0.800. On my current project (interlaced DV and HDV to 1080p) I'm actually typically using BOTH LSFMod (similar to the LimitedSharpenFaster sharpening in RenderPlus) of 300 to 500, AND Vegas sharpening of 0.500 to 0.700.

Also...

3. The RenderPlus resizer crops 56px from the top and bottom. However this gives a slightly vertically squashed result. I think the reason for this may be rooted in the non-square pixels of NTSC video. I calculated that 60px would be correct, but this gave me narrow black bars top and bottom. So actually I think it's better to resize to 1280x960 then crop 120px from the top and bottom either in AviSynth or with VEGAS Pan/Crop.

I did make another script that you can save as .avs and load into RenderPlus as a custom script to try:

SetMemoryMax(4096)
SetFilterMTMode("DEFAULT_MT_MODE", 2)
AviSource("d:\fs.avi")
AssumeBFF()
ConvertToYV12(interlaced=true)
QTGMC(Preset="Slower")
LanczosResize(1280,960)
Crop(0,120,-0,-120)
LimitedSharpenFaster(strength=800)
Prefetch(4)

But unfortunately the rendered file is flagged as interlaced, and I can't work out why. @wwaag?

MikeLV wrote on 3/10/2020, 11:20 AM

After some testing, I think it will probably better to pillarbox instead. And I think the sharpening isn't good, at least with the resize. Is there a way to use the script but have it pillar box instead? So far, the better result is with HOS settings, without the script

NickHope wrote on 3/10/2020, 12:08 PM

As said previously, you may well not have to pillarbox your render to 16:9 at all (like this 4:3 720p video based on PAL DV). Try uploading a 960x720 render with no black side bars. Script for that:

SetMemoryMax(4096)
SetFilterMTMode("DEFAULT_MT_MODE", 2)
AviSource("d:\fs.avi")
AssumeBFF()
ConvertToYV12(interlaced=true)
QTGMC(Preset="Slower")
LanczosResize(960,720)
LimitedSharpenFaster(strength=800)
Prefetch(4)

If you need a script that renders with black side bars, I could look into that tomorrow. I think the RenderPlus resizer offers that anyway.

MikeLV wrote on 12/2/2020, 1:50 PM

Did something change is HOS? When I go to the resize/crop settings dialog, I can't make the pillarbox bars appear on the sides like it used to? Or is there a better way to go from 480i to 720p now? @wwaag Or maybe I'm supposed to be using a script, can't recall.

fifonik wrote on 12/2/2020, 10:09 PM

Do you have source sample somewhere?

Camcorder: Panasonic X1500 + Panasonic X920 + GoPro Hero 11 Black

Desktop: MB: MSI B650P, CPU: AMD Ryzen 9700X, RAM: G'Skill 32 GB DDR5@6000, Graphics card: MSI RX6600 8GB, SSD: Samsung 970 Evo+ 1TB (NVMe, OS), HDD WD 4TB, HDD Toshiba 4TB, OS: Windows 10 Pro 22H2

NLE: Vegas Pro [Edit] 11, 12, 13, 15, 17, 18, 19, 22

Author of FFMetrics and FFBitrateViewer

MikeLV wrote on 12/2/2020, 10:22 PM

Do you have source sample somewhere?

Do you mean the original footage?

wwaag wrote on 12/2/2020, 10:29 PM

@MikeLV

Nothing has changed in HOS regarding the resize/crop settings dialog in R+. Exactly, which tool are you trying to use? If you are using R+, you can simply use the dialog. If you are using AviDub, you must use a script like the one posted by NickHope back in March just a few posts up. If you need help, send me a PM with the specifics.

AKA the HappyOtter at https://tools4vegas.com/. System 1: Intel i7-8700k with HD 630 graphics plus an Nvidia RTX4070 graphics card. System 2: Intel i7-3770k with HD 4000 graphics plus an AMD RX550 graphics card. System 3: Laptop. Dell Inspiron Plus 16. Intel i7-11800H, Intel Graphics. Current cameras include Panasonic FZ2500, GoPro Hero11 and Hero8 Black plus a myriad of smartPhone, pocket cameras, video cameras and film cameras going back to the original Nikon S.

MikeLV wrote on 12/2/2020, 10:33 PM

Using R+, when I go to the crop settings window, and click on pillar box, and click the preview button after choosing the new frame size, it does not show the black bars on the left and right of the video like it did before.

wwaag wrote on 12/2/2020, 10:41 PM

@MikeLV

Will check.

AKA the HappyOtter at https://tools4vegas.com/. System 1: Intel i7-8700k with HD 630 graphics plus an Nvidia RTX4070 graphics card. System 2: Intel i7-3770k with HD 4000 graphics plus an AMD RX550 graphics card. System 3: Laptop. Dell Inspiron Plus 16. Intel i7-11800H, Intel Graphics. Current cameras include Panasonic FZ2500, GoPro Hero11 and Hero8 Black plus a myriad of smartPhone, pocket cameras, video cameras and film cameras going back to the original Nikon S.

wwaag wrote on 12/2/2020, 10:50 PM

I can confirm that something is messed up. Will fix tomorrow.

AKA the HappyOtter at https://tools4vegas.com/. System 1: Intel i7-8700k with HD 630 graphics plus an Nvidia RTX4070 graphics card. System 2: Intel i7-3770k with HD 4000 graphics plus an AMD RX550 graphics card. System 3: Laptop. Dell Inspiron Plus 16. Intel i7-11800H, Intel Graphics. Current cameras include Panasonic FZ2500, GoPro Hero11 and Hero8 Black plus a myriad of smartPhone, pocket cameras, video cameras and film cameras going back to the original Nikon S.

MikeLV wrote on 12/2/2020, 10:51 PM

👍👍

fifonik wrote on 12/3/2020, 1:31 AM

Do you mean the original footage?

Yes (short fragment), so I can try my avisynth script.

Camcorder: Panasonic X1500 + Panasonic X920 + GoPro Hero 11 Black

Desktop: MB: MSI B650P, CPU: AMD Ryzen 9700X, RAM: G'Skill 32 GB DDR5@6000, Graphics card: MSI RX6600 8GB, SSD: Samsung 970 Evo+ 1TB (NVMe, OS), HDD WD 4TB, HDD Toshiba 4TB, OS: Windows 10 Pro 22H2

NLE: Vegas Pro [Edit] 11, 12, 13, 15, 17, 18, 19, 22

Author of FFMetrics and FFBitrateViewer

MikeLV wrote on 12/3/2020, 9:56 AM

Do you mean the original footage?

Yes (short fragment), so I can try my avisynth script.

I sent a private message to you @fifonik

wwaag wrote on 12/3/2020, 12:38 PM

See this. https://www.vegascreativesoftware.info/us/forum/happy-otter-scripts-for-vegas-pro--113922/?page=32#ca782183

AKA the HappyOtter at https://tools4vegas.com/. System 1: Intel i7-8700k with HD 630 graphics plus an Nvidia RTX4070 graphics card. System 2: Intel i7-3770k with HD 4000 graphics plus an AMD RX550 graphics card. System 3: Laptop. Dell Inspiron Plus 16. Intel i7-11800H, Intel Graphics. Current cameras include Panasonic FZ2500, GoPro Hero11 and Hero8 Black plus a myriad of smartPhone, pocket cameras, video cameras and film cameras going back to the original Nikon S.